Xerostomia, Xerogenic Medications and Food Avoidances in Selected Geriatric Groups W. J. Loesche, DMD, PhD,'tJ. Bromberg, PhD,# M. S. Terpenning, MD,+$W. A. Bretz, DrPH,' B. L. Dominguez, MS,'$ N. S. Grossman, MA,' and S. E. Langmore, PhD'$

CONCLUSION Xerostomia apparently affects the ability to OBJECTIVE: To study the relationship between complaints chew and start a swallow. This leads to avoidance of certain of xerostomia and salivary performance and food avoidances foods, which raises the possibility that xerostomia could in four geriatric groups chosen to reflect a broad spectrum of contribute to undernutrition in older persons. The topically individuals along the health-disease continuum. To deter- applied ipratropium and triamcinolone and the systemic mine whether xerogenic medications taken by these individ- agents amitriptyline, oxybutynin and triazolam could be uals could be associated with either complaints of xerostomia statistically associated with one or more complaints of xer- or with food avoidances. ostomia. J Am Geriatr SOC43:401407, 190. DESIGN. Cross-sectional survey. Clinical examinations and interview. SETTING: A VA dental clinic and a retirement home. oor dental health, as measured by the number of missing Pteeth and/or the number of decayed or periodontally PATIENTS AND OTHER PARTICIPANTS: Subjects were diseased teeth, can affect dietary choices with resultant ad- 529 individuals older than 56 years of age, living both in verse nutritional consequences.' This problem is potentially institutions and in the community (average age 70 years). important in older persons, where the prevalence of oral/ Two hundred eight persons were recruited from a VA Dental dental problems is Sullivan and colleagues have Clinic, 114 from a residential retirement home, and 132 from shown that poor oral hygiene,' xerostomia, and the inability a nursingllong-term-care facility; 75 were from an acute care to chew, among other oral conditions, were the best predic- ward and had a diagnosis of a cerebral vascular accident or tors of significant involuntary weight loss in the previous year other neurological condition. among a population of frail older persons. In related studies, OUTCOME MEASURES Prevalence of xerostomia, dental weight loss in the year before admission was one of the morbidity, salivary flow, and food avoidances. strongest predictors of subsequent in-hospital morbidity and RESULTS: About 72% of the subjects experienced xerosto- mortality,* and involuntary weight loss was a strong predic- mia sometime during the day. Stimulated salivary flow was tor of 1-year postdischarge rn~rtality.~ found to be significantly lower in individuals who com- Xerostomia is a common complaint of older people and plained of xerostomia than in those who did not. Fifty-five is usually associated with reduced salivary flo~.'~-'~While percent of participants reported using one or more xerogenic some cases of xerostomia are related to auto-immune dis- medications, with an 86% prevalence in the nursingllong- eases such as Sjogren's ~yndrome,'~and others are secondary term-care facility. Individuals with xerostomia had difficulty to irradiation treatment of oral neoplasm^,'^ the majority of in chewing and in starting a swallow and were significantly the cases appear to be related to the usage of certain medica- more likely to avoid crunchy foods such as vegetables, dry tions.'5-'8 More than 400 medications have been cited as foods such as bread, and sticky foods such as peanut butter in xerostomia-inducing or xerogenic drugs," but with the ex- their diets. Specific medications such as the inhalants iprat- ception of tricyclic antidepressants,20 anticholinergic agents, ropium and triamcinolone and the systemic agents oxybuty- certain psychotrophic medications and antihistamines, and nin and triazolam could be statistically associated with xer- possibly diuretics:' most of the evidence linking a particular ostomia and/or low salivary flow, and/or specific food medication with xerostomia is anecdotal. In addition, several avoidances. studies have shown that the number of medications used by the individual per day, regardless of whether they are xero- genic, can be associated with xer~~t~mia.'~~~~~~~ The aim of the present investigation was to study the From the 'University of Michigan School of Dentistry, tuniversity of Michigan School of Medicine, $University of Michigan Department of Eiostatistics School prevalence of complaints of xerostomia in four geriatric of Public Health and SAnn Arbor Veterans Affairs Medical Center Ann Arbor, groups chosen to reflect a broad spectrum of individuals Michigan. along the health-disease continuum (Loesche et al., unpub- This investigation is supported by USPH Research Grant DE-09142 from the lished data, 1994). We then determined whether xerostomia National Institute of Dental Research, and IP30-AG-008808 from the National Institutes of Health, Bethesda, Maryland. could be associated with reduced salivary performance and Address correspondence to Walter J. Loesche, DMD, PhD, University of Michi- with the avoidance of various types of food. We also investi- gan School of Dentistry, 1011 N. University, Ann Arbor, Michigan 48109-1078. gated which xerogenic medications taken by these individuals

JAGS 43:401407,1995 0 1995 by the American Geriatrics Society 0002-8614/95/$3.50 402 LOESCHE ET AL. APRIL 1995-VOL. 43, NO. 4 JAGS could be associated with either complaints of xerostomia or their dental examination. Only those subjects who answered with avoidance of certain types of food. yes to the xerostomia questions on both occasions were considered as having xerostomia for our analysis. MATERIALS AND METHODS For VA participants, the reported current use of medica- Sample Population tions was compared with the computerized pharmacy records maintained by the Ann Arbor VA Medical Center. For the Subjectdpatients were chosen with the intention of hav- non-VA participants the individual was asked to bring all ing groups with different medical backgrounds. All partici- medications currently being taken, to the interview, and only pants were at least 56 years of age (average age was 70 years) these medications were recorded on our forms. Medications and signed a consent form approved by an institutional were identified as xerogenic using the PDR Drug Interactions/ review board at both the University of Michigan School of Side Effectdndications software program version 2.0 (Physi- Dentistry and the Ann Arbor Veterans Affairs Medical Cen- cian’s Desk Reference, Medical Economical Data Inc., ter. The characteristics of these populations have been de- Montvale NJ). scribed elsewhere and will be briefly summarized (Loescheet al., unpublished data, 1994). Dental Examination All VA patients were examined by the same clinician/ Independent Living Groups dental assistant team at the VA dental facility. All non-VA This group consisted of 208 individuals recruited from patients were examined by a second cliniciatddental hygienist the VA Dental Clinic and 114 individuals recruited from team using the dental facility at the Glacier Hills Retirement either the Glacier Hills Retirement Center, Ann Arbor, Mich- Center. Both teams were initially calibrated by the same igan or from the Turner Geriatric Clinic at the University of individual (WB) and were recalibrated semiannually. The Michigan Medical Center, Ann Arbor, Michigan. Ninety- number of teeth that were present and the number of de- four percent of the VA subjects were male, and 94% were cayed/missing/filled teeth and surfaces (DMF teeth, DMF white. Their average age was 67.8 years 5 5.1 years. Thirty surfaces) were recorded on standardized forms.6 The pres- percent of the non-VA subjects were male and 98% were ence of any fixed or removable prostheses, dentures or im- white. Their average age was 78 2 8.5 years. Because of these plants was recorded. Edentulousness was defined for our differences in age and gender, two separate independent analysis as the absence of natural teeth and did not take into living groups were established (Loesche et al., unpublished account whether dentures were present. Approximately 90% data, 1994). of the edentulous individuals had dentures available. Dependent Living Groups Saliva The nursing home group consisted of 122 residents of the Whole saliva was stimulated by swabbing the tongue long-term-care facility of the VA Hospital and 10 residents of with a 2% citric acid solution three times at 30-second the Glacier Hills Nursing Home. Ninety-two percent of the interval^.'^ During the next minute, the patients swallowed Nursing Home group were males, and 89% were white. the first flow of saliva containing the citric acid. Thereafter Their average age was 70 2 7.2 years. the stimulated saliva was collected over a 5-minute period by The hospitalized group consisted of 75 VA patients, who asking the patients to tilt their head forward and to spit their had been admitted to the acute care unit with a diagnosis of a saliva into a graduated, preweighed, conical tube. If the cerebral vascular accident or other neurological condition. patient was unable to provide about 1 mL of saliva after 5 About 90% of these individuals had been in an independent minutes, hdshe was given a default value of <0.2 mUmin. living status before their admission. One hundred percent of All VA Hospitalized patients had their saliva collected by a the Hospitalized group were males, and 81% were white. suction procedure:’ as it was not possible for them to spit. Their average age was 67 2 6.0 years. The amount of saliva, as determined gravimetrically, the length of collection, and the time of day of the collection were Questionnaire also recorded. The minor salivary gland output was mea- All participants were interviewed by trained interviewers sured as the degree of wetness present in specially prepared who used a structured questionnaire to elicit information standardized cellulose stripsF6 These strips were individually about demographic characteristics, medical and dental his- placed on the hard palate, the buccal mucosa, and the inside tory, use of medications, complaints of xero~tomia,2~and lip and were allowed to absorb moisture for 30 seconds. The various food avoidances (questionnaire available upon re- wetness was measured using a PeriotronTM instrument that quest). The questionnaire was pretested for comprehension in was calibrated each day against a known volume of serum. a subsample of older individuals. In a few cases, mostly in the The findings were reported as FUmin for each location, and Hospitalized group, the information was obtained with the the sum of the three intraoral locations was reported as the assistance of a relative. The interview lasted from 1 to 1’12 minor salivary gland output. hours, and provided detailed information relative to the subject’s eating habits, food and liquid preferences, avoid- Statistical Analysis ance of food types, oral hygiene habits, complaints of a dry If the attained probability of the overall comparisons for mouth (xerostomia), swallowing complaints, medical his- any given parameter was less than P = .05, then pair-wise tory, and usage of prescription medications. We asked the comparisons were performed using Fisher’s Exact test for subjects questions regarding the time of the day in which they categorical variables at two levels. The large number of experienced a dry mouth and then questions regarding the comparisons being made caused us to accept a P < .01 as amount of saliva they perceived. The questions on xerosto- being significant for the pairwise comparisons. Significance in mia were again asked by the dentist when the subjects had those parts of Tables 2 to 4 that illustrate discrete 2x2 JAGS APRIL 1995-VOL. 43, NO. 4 ORAUDENTAL HEALTH IN GERIATRIC INDIVIDUALS 403 associations was obtained using Fisher’s Exact Test. The their teeth (Table 2). These individuals were significantly underlying differences between groups for the continuous more likely to report that food sticks in their throat and that variables in Tables 3 and 4 were found using the nonpara- they had difficulty in chewing, in starting a swallow, and in metric Wilcoxon rank-sum test. This nonparametric test was swallowing (Table 2). They were 6.4 times more likely to used because of the lack of a normal distribution in the report that they had too little saliva, and their mouths were parameters under investigation. 5.2 times more likely to be dry when eating. Perhaps to compensate for this dryness, they were 2.1 times more likely RESULTS to report that they sip liquids to help them swallow dry foods. There were no significant differences between the four These findings suggested that an individual with a com- groups when complaints of xerostomia were compared plaint of xerostomia might have problems eating because of (Table 1).About 72% of the subjects experienced xerostomia these difficulties associated with chewing and swallowing. at some time during the day, with about 46% reporting this We asked all subjects whether there were certain foods that complaint either in the morning or at night. Complaints in the s/he avoided in their diets. Thirty-one percent of the individ- evening were significantly higher for the Hospitalized pa- uals with a complaint of “xerostomia in the morning” stated tients. Only 13% of the participants had complaints when that they avoided one or more food types, a prevalence that eating (Table 1).There was no effect of age on the complaint was significantly higher than the response rate of 18% re- of xerostomia; both complainers and noncomplainers aver- ported by people without the complaint (Table 2). The indi- aged 70.3 to 70.7 years in age. There was no relationship viduals who complained of xerostomia, relative to the non- between edentulousness and xerostomia, as 30% of both the xerostomic individuals, were significantly more likely to complainers and noncomplainers were edentulous. There avoid crunchy foods such as carrots and sticky foods such as was no obvious effect of gender because the mostly female, peanut butter. There was also a tendency to avoid dry foods non-VA subjects did not have a significantly lower level of such as bread. complaints than did the male VA subjects. Because of these We determined whether the complaint of xerostomia similarities, the four groups were combined for the subse- could be associated with the usage of putative xerogenic quent analysis of xerostomia. medications. About 46 to 48% of the Independent Living A complaint of xerostomia at any time of the day was associated with a significant reduction in both stimulated subjects, 55% of the Hospitalized patients, but 86% of the Nursing Home patients used one or more xerogenic medica- salivary flow (0.6920.35 mYmin vs 0.58 5 0.31 mumin, P < .001 Wilcoxon rank sum test), and in minor salivary tions. The higher frequency of usage in the Nursing Home gland output (5.8 2 2.3 pUmin vs 5.3 2 2.5 pUmin, patients was significant when compared with the other P = .03). A complaint of xerostomia while eating was asso- groups, i.e., P < .001. However, despite this higher usage of ciated with the greatest reduction in stimulated salivary flow putative xerogenic medications in the nursing home, these (0.63 -t 0.33 mumin vs 0.41 -t 0.22 mumin), but with the individuals did not have a significantly higher prevalence of smallest reduction in minor salivary gland output (5.5 pU xerostomic complaints (Table 1).The individuals who used min vs 5.1 pUmin). We chose “xerostomia in the morning” xerogenic medications were 2.1 times more likely to have a as the complaint of reference for all subsequent analysis complaint of xerostomia in the morning, than individuals because this complaint was the most frequently reported and who did not use xerogenic medications (P < 0.001). These because the magnitude of difference between complainers individuals also had significantly less stimulated salivary and noncomplainers in regard to stimulated salivary flow flow, but there was no apparent effect of the usage of (0.12 mumin), and minor salivary gland output (0.9 pU xerogenic medications on the minor salivary gland output min), was the greatest in individuals with this complaint. (Table 3). Individuals who used xerogenic medications were We determined whether this complaint of xerostomia in 2.5 times more likely to be edentulous and, among the den- the morning could be associated with oral problems relating tate individuals, were 2.7 times more likely to have fewer to dryness. The individuals who complained of xerostomia than 10 teeth, compared with those individuals who did not were 14.1 times more likely to state that their cheeks stuck to use xerogenic medications (Table 3).

Table 1. Frequency of Complaints of Xerostomia Among the Various Geriatric Groups

Independent Living Dependent Living All Subjects non-VA VA Nursing Home Hospitalized Complaint of Xerostomia n = 529 n = 114 n = 208 n = 132 n = 75

Any timet 72%* 68%* 72%‘ 71 %* 82%* In morning 46 48 40 51 54 In afternoqn 31 27 28 33 39 In evening 32 27 26 35 49* At night 45 43 42 49 50 When eating 13 8 13 15 18

‘Proportion of individuals in each group who gave a positive response for complaint. ‘Any time would include individuals who had one or more complaints related to the five temporal periods listed. ‘Overall difference is significant, P = 0.004, due primarily to high proportion of complaints in hospitalization patients relative to both independent living groups. 404 LOESCHEETAL. APRIL 1995-VOL. 43, NO. 4 JAGS

Table 2. Relationship Between Complaints of Xerostomia in the Morning and Complaints Relative to Oral Dryness and Eating and Salivary Hypofunctions

Complaints of Xerostomia in Morning Odds Ratio No (281)* Yes (242) Significance* (95% CI) Oral dryness Cheeks stick to teeth 1.1 %* 13.5% <0.001 14.1 (4.2,46.7) Food stick in throat 17.3 34.2 <0.001 2.5 (1.6,3.7) Difficulty in chewing 12.2 27.2 <0.001 2.7 (1.7,4.3) starting a swallow 8.7 18.9 <0.001 2.4 (1.4,4.2) in swallowing5 19.0 34.8 <0.001 2.3 (1.5, 3.4) Subjective measures of salivary hypofunction Sip liquids5 26.8 43.9 <0.001 2.1 (1.5, 3.1) Mouth dry when eating5 5.1 21.8 <0.001 5.2 (2.8,9.7) Is saliva too little5 7.9 35.6 <0.001 6.4 (3.9, 10.7) Types of foods avoided Any food 17.7%* 31.5%* <0.001 2.1 (1.4,3.2) Chewy (beef)" 12.3 17.3 NS Crunchy (carrot) 9.4 17.7 0.006 2.1 (1.2,3.5) Dry (bread) 5.4 10.0 0.065 2.0 (1 .O, 3.8) Crumbly (Cake) 3.6 6.3 NS Sticky (peanut butter) 2.9 8.0 0.016 2.9 (1.2,6.8)

'Number of subjects in each response group. 'Significance - Fisher's Exact test - 2 tail. *Percentage of subjects who responded yes to question. SQuestions related to salivary hypofunction. See ref. 25. "Example of food type explained to patient when asking question in food avoidance.

Table 3. Relationship Between Usage of Xerogenic Medications and Salivary Function and Dental Status

Simulated Minor Salivary Usage of Xerogenic No. of Salivary Flow Gland Output % % Dentate Medications Subjects rnVmin pUmin. Edentulous 2 10 teeth

No 222 0.64 5 0.34 5.2 2 2.3 20 72% Yes 307 0.58 2 0.32 5.5 2 2.6 37 49% Significance P = 0.025" NS P < .001t P < .001+ Odds Ratio (95% 2.5 (1.6,3.7) 2.7 (1.8,4.0) confidence interval)

'Wilcoxon rank-sum test. 'Fisher's Exact Test.

We next examined the relationship between the reported subjects), and thiethylperazine (n = 9 subjects) could not be usage of individual putative xerogenic medications with the associated with any complaint of xerostomia or with reduced complaints of xerostomia and with the measured salivary salivary flow. However, with these latter medications the parameters. The 20 xerogenic medications used most fre- number of subjects may have been too small to draw any quently by these subjects were examined for their association conclusions. with any complaint of xerostomia or with any problems of Seven medications, ipratropium (Atrovent), triamcin- salivary gland performance. The three most frequently pre- olone (Azmacort), oxybutynin (Ditropan), amitriptyline scribed putative xerogenic medications, diltiazem (n = 55 (Elavil), triazolam (Halcion), sucralfate (Carafate), and ibu- subjects), diphenhydramine (n = 49 subjects), and triam- profen (Motrin)could be significantly associated with com- terenehydrochlorothiazide (n = 39 subjects) could not be plaints of xerostomia at one or more times during the day associated with any complaint of xerostomia or with any (Table 4). The individuals who used ipratropium were 15 form of reduced salivary gland performance. Likewise, the times more likely to report any complaint of xerostomia than usage of verapamil (n = 20 subjects), timolol (n = 21 sub- were individuals who did not use this bronchodilator. De- jects), hydroxyzine (n = 12 subjects), clonidine (n = 11 spite this strong association with complaints of xerostomia, subjects), metoprolol (n = 11 subjects), guaifenesin (n = 10 ipratropium was not associated with a reduced stimulated JAGS APRIL 1995-VOL. 43, NO. 4 ORAyDENTAL HEALTH IN GERIATRIC INDIVIDUALS 405

Table 4. Relationship Between Usage of Certain Xerogenic Medications and Complaints of Xerostomia, and Salivary Parameters

Xerostomia Salivary Parameters Any Time In Morning At Night/in Dry while Too Little Stimulate MSGO evening eating (Subjective) d

Xerogenic* ~~ Medication (N) ORt Sig(P)+ OR Sig(P) OR Sig(P) OR Sig(P) OR Sig(P) Sig(P)ll Sig(P)"

lpratropium (36) 14.9 <.001 2.1 .05 2.9 <.003 3.2 .006 2.9 .005 .03$ Triamcinolone (1 3) 6.4 .009 4.1 .03 3.4 .06 6.2 .002 Oxybutynin (1 8) 4.1 .01 3.2 .03 Triazolam (38) 2.2 .03 .01 Amitriptyline (1 8) 2.8 .04 4.9 .001 .02 .08 Fluoxetinne (9) .02 Sucralfate (27) 2.9 .03 .03 Ibuprofen (37) 2.6 .03 Transderm-nitro (1 9) .03

*Number of subjects using medication. +Sig = significance using Fisher's Exact Test; OR = Odds Ratio. All values are significant in that the 95% confidence interval does not include 1.0. 'Higher in group taking ipratropium. SMSGO = minor salivary gland output. "Differencesbetween subjects using and not using medication significant using Wilcoxon rank sum test. salivary flow, and there was actually a significant increase in flect the fact that participants were asked multiple questions the minor salivary gland output (Table 4). The patients who about the time of day in which they experienced mouth used ipratropium were more likely to be edentulous, and if dryness, were queried about xerostomia on two separate dentate, to have fewer than 10 teeth (data not shown). occasions, and that the xerostomia questions were positioned Individuals who used the other inhalant on this list, triamci- after those questions related to thirst. Our positive xerosto- nolone, were 6 times more likely to report a complaint of mia responses would also have been increased by the inclu- xerostomia in the morning and to state that they had too little sion of hospitalized patients with a swallowing problem. saliva (Table 4). There was no observed reduction in either Eighty-two percent of these patients reported having a com- stimulated salivary flow or in the minor salivary gland out- plaint of xerostomia (Table l), but as they comprised only put. The subjects who used oxybutynin, triazolam, and ami- 14% of our total population, they alone, could not account triptyline were significantly more likely to complain of xer- for the high prevalence of xerostomia noted. ostomia in the evening or at night. The usage of sucralfate Xerostomia has been associated with various oraYdental and ibuprofen were associated with dryness only when eat- problems, such as an increased prevalence of de~ay,'~~~' ing. edentulousness?' and salivary hypofun~tion.'~~~~We ob- The use of triazolam, amitriptyline, fluoxetine, and su- served similar associations in that our subjects with com- cralfate was associated with a significantly reduced stimu- plaints of xerostomia had significantly more decayed teeth, lated salivary flow, but with the exception of amitriptyline, (unpublished data), and reduced stimulated whole saliva and there was no effect on the minor salivary gland output (Table minor salivary gland output. Some investigators have found 4).Although none of the nine individuals who used fluoxetine no relationship between complaints of xerostomia and stim- had a complaint of xerostomia, there was a significantly ulated salivary flow.23,24y29Several investigators have sug- lower stimulated salivary flow in these patients. Transderm- gested that resting whole saliva prod~~tion,Z~~~~~~'which has nitro and, to a lesser extent, amitriptyline were associated a large contribution from the submandibular gland and lesser with lower minor salivary gland output (Table 4). amounts from the parotid, sublingual and the various minor We next determined whether any of these 20 putative salivary glands:0J' is the most reliable objective correlate of xerogenic medications could be associated with any reported xerostomia. Resting whole saliva is difficult to measure, es- food avoidances. Individuals using ipratropium were signifi- pecially in the sick and in older persons, and because of this is cantly more likely to avoid one or more types of food (OR = often not included in investigations of ~erostomia.~~~~~We 2.28, P = .04), especially chewy (meats) (OR = 3.2, P = have used the minor salivary gland output as a surrogate for .005)and crunchy (carrots) (OR = 2.8, P = .02) types of resting whole saliva because its collection is simple and re- food. The individuals taking amitriptyline were 7 times more producible.26Although the minor salivary gland output may likely to avoid chewy foods (P = .007), and those using only comprise up to 8% of the resting salivary flo~:~,~'we diphenhydramine were 2.4 times more likely to avoid found that it was significantly associated with complaints of crunchy foods (P = .02). xerostomia. This suggests that the minor salivary gland out- put may serve as a surrogate for resting whole saliva in DISCUSSION studies involving xerostomia. The 72% prevalence of complaints of xerostomia at any Complaints of xerostomia in our subjects could be sig- time during the day reported by the participants in this nificantly associated with a positive response to a series of investigation (Table 1) is higher than the prevalences of 10 to questions that have been shown to significantly correlate with 55% found by other investigat~rs.'~-'~~'~~'' This could re- hypofunction of the salivary glands,Z4 (Table 2). The xeros- 406 LOESCHE ET AL. APRIL 1995-VOL. 43, NO. 4 JAGS tomic individuals were 6.4 and 5.2 times more likely to give a When considering the relationship between drugs and positive response to questions about having too little saliva salivary glands, it is necessary to recognize the difference and having a dry mouth when eating, but they were 14 times between medications that cause the sensation of dryness, but more likely to state that their cheeks stuck to their teeth or which may have no effect on the salivary glands, and those dentures (Table 2). As such, these three questions might medications that can actually inhibit or alter salivary secre- reliably identify individuals with salivary gland hypofunc- tions21 Other investigators have noted that oral dryness does tion. not reliably indicate decreased salivary o~tput.~'?~'*~~With Individuals with xerostomia in the morning reported ipratropium bromide and triamcinolone acetonide, this dif- having difficulty in chewing, in starting a swallow, and in ference was clearly observed as these topical agents caused swallowing (Table 2). This could influence their choice of dryness of the oral surfaces, yet had no discernible effect on foods, and, indeed, individuals with a complaint of xerosto- salivary flow. In this case, the absence of an effect probably mia were significantly more likely to avoid crunchy foods reflected that these topically delivered agents did not achieve such as carrots and sticky foods such as peanut butter (Table systemic levels that could influence salivary secretion. 2). Other investigators have shown that xerostomia can be It was of interest to determine whether any of the twenty associated with undernutrition among both institutionalized most frequently used medications could be associated with a and independent-living older persons 34 and that xerostomia history of food avoidances. Ipratropium could be associated is one of the common oral conditions associated with signif- with the avoidance of chewy and crunchy foods. This obser- icant involuntary weight loss among frail older people.' Our vation suggests that this topically applied agent was causing observations on food avoidances, albeit indirect measures of either a degree of dryness or a sense of dryness that made dietary intake, support the hypothesis that a complaint of certain foods uncomfortable to eat, despite the fact that the xerostomia may contribute to undernutrition in older per- salivary flow was not obviously reduced. Individuals who sons. used amitriptyline were 7 times more likely to avoid chewy These adverse medical outcomes associated with com- foods such as meat. Diphenhydramine, a putative xerogenic plaints of xerostomia warrant efforts to prevent and treat this medication, for which we could find no association with a condition. More than 400 medications are believed to be complaint of xerostomia or of a reduced salivary flow, was, xerogenic," which suggests that one approach to prevent or however, associated with an avoidance of crunchy type to reduce xerostomia would be to substitute a nonxerogenic foods. None of the medications associated with a reduction in medication for a xerogenic medication whenever the medical salivary flow, with the exception of amitriptyline, could be condition permits such an option. associated with the avoidance of any types of food, suggest- We looked at the relationship between the 20 most ing that the sense of dryness, rather than a lack of saliva, was frequently used medications and the various oravdental pa- the most important consideration in the subject's food pref- rameters under investigation. The confounding effect of the erences. In the case of amitriptyline, there were also com- use of more than one xerogenic medication, or of varying plaints of xerostomia, reinforcing the suggestion that xeros- dosages of the medication and patient compliance, were not tomia per se was the main determinant of food avoidances. addressed. Nine medications could be associated significantly The present findings confirm other studies that show that with either a complaint of xerostomia or with diminished complaints of xerostomia are common among older per- salivary gland performance. Triazolam, amitriptyline, and sons10,11,15-18,21 and that these complaints can be associated sucralfate, which could be significantly associated with both with the use of one or more medications that are considered a complaint of xerostomia and with a reduction in stimulated to be xer~genic.'~.'1315-18-21 Specific medications were iden- salivary flow, were the most xerogenic of these medications. tified that appeared more likely to be associated with xeros- Of these, amitriptyline and triazolam, which have been tomia than others. Of these, the topically applied iprat- grouped by several authors with antipsychotic, hypnotic, and ropium and triamcinolone, both inhalants used for bronchial psychotropic medications, have frequently been associated dilation, and the systemic agents oxybutynin, triazolam, and with xer~~tomia.~~~~~~~~~~'-~~~~~~~~~~~~~~There is apparently amitriptyline, could be statistically associated with one or little data that associates sucralfate with decreased salivary more complaints of xerostomia. Of these, only triazolam and gland performance.2' amitriptyline could be associated with a reduced salivary Ipratropium bromide and triamcinolone acetonide, used flow. as inhalants in the treatment of emphysema and asthma, were significantly associated with complaints of xerostomia and ACKNOWLEDGMENTS with the patients own assessment that they had too little Carol Gerlach assisted in the preparation of this manu- saliva (Table 4). However, neither agent significantly affected script. Gary Hildebrandt and Katherine Curry assisted in the stimulated salivary flow and perhaps, because of the mucosal collection of dental data. Bonnie Hartsuff assisted in the dryness, there was a compensatory increase in the minor statistical analysis. salivary gland output when ipratropium was used (Table 4). Oxybutynin and ibuprofen could be associated with com- REFERENCES 1. Geissler CA, Bates JF. The nutritional effects of tooth loss. Am J Clin Nutr plaints of xerostomia but not with any decrease in salivary 1984;39:478 -489. gland performance. Fluoxetine, although used by only nine 2. Shatenstein B. Tooth loss, mastication, and oral health in the elderly. J Am subjects, could be associated with a reduction in stimulated Diet Assoc 1986;86:86-87. salivary flow. Previously we had found that fluoxetine signif- 3. Howath CC. Chewing difficulty and dietary intake in the elderly. J Nutr Eld- icantly reduced stimulated salivary flow in young, bulimic erly 1989;9:17-24. 4. Hatton ER, Gogan CM, Hatron MN. Common oral conditions in the eld- women.37 This would suggest that the substitution of fluox- erly. Am Fam Physician 1989;40:149-162. etine for any of the tricyclic antidepressants may not improve 5. Neissen LC, Jones JA. Oral health changes in the elderly: Their relationship salivary flow. to nutrition. Postgrad Med 1984;75:231-237. IAGS APRIL 1995-VOL. 43. NO. 4 ORAUDENTAL HEALTH IN GERIATRIC INDIVIDUALS 407

6. Miller AJ, Brunelle JA, Carlos JP et al. Oral health of United States adults. 22. Sreebny LM, Valdini A, Yu A. Xerostomia. Part 11: Relationship to nonoral Bethesda, MD: National Institute of Health. DHHS Pub No.87-28681987. symptoms, drugs, and disease. Oral Surg Oral Med Oral Pathol 7. Sullivan DH, Martin W, Flaxman N et al. Oral health problems and invol- 1989;68:419-427. untary weight loss in a population of frail elderly. J Am Ceriatr Soc 23. Thorselius I, Emilson CG, Osterberg T. Salivary conditions and drug con- 1993;41:725-73 1. sumption in older age groups of elderly Swedish individuals. Gerodontics 8. Sullivan DH, Patch GA, Walls RC et al. Impact of nutritional status on mor- 1988;4:66-70. bidity and mortality in a select population of geriatric patients. Am J Clin 24. Fox PC, Buxh KA, Baum BJ. Subjective reports of xerostomia and objective Nutr 1990;51:749-758. measures of salivary gland performance. J Am Dent Assoc 1987;115:581- 9. Sullivan DH, Walls RC, Lipschitz DA. Protein-energy undernutrition and the 584. 25. Fox PC, van der Ven PF, Sonies BS et al. Xerostomia: Evaluation of a symp- risk of mortality within 1 y of hospital discharge in a select population of tom with increasing significance. J Am Dent Assoc 1985;110:519-525. geriatric patients. Am J Clin Nun 199153599-605. 26. Shern RJ, Fox PC, Cain JL et al. A method for measuring the flow of saliva 10. Ben-Aryeh H, Miron D, Berdicevsky I et al. Xerostomia in the elderly: Preva- from the minor salivary glands. J Dent Res 1990;69:1146-1149. lence, diagnosis, complications and treatment. Gerodontology 1985;4:77- 27. Papas AS, Joshi A, MacDonald SL et al. Caries prevalence in xerostomic in- 82. dividuals. J Can Dent Assoc 1993;59171-174. 11. Osterberg T, Birkhed D, Johansson C et al. Longitudinal study of stimulated 28. Sreebny LM, Valdini A. Xerostomia. Part 1. Relationship to other oral symp- whole saliva in an elderly population. Scand J Dent Res 1992;100:340-345. toms and salivary gland hypofunction. Oral Surg Oral Med Oral Pathol 12. Sreebny LM, Valdini .A Xerostomia. A neglected symptom. Arch Intern 1988;66:451-458. Med 1987;147:1333-1337. 29. Atkinson JC, Shiroky JB, Macynski A et al. The effects of furosemide on sali- 13. Mandel ID. Sialochemistry in diseases and clinical situations affecting sali- vary secretions. Cerodontology 1989;16:23-28. vary glands. CRC Crit Rev Clin Lab Sci 1980;12:321-366. 30. Schneyer LM. Source of resting total mixed saliva of man. J Appl Physiol 14. Dreizen S, Brown LR, Handler S et al. Radiation-induced xerostomia in can- 1956;979- 8 1. cer patients. Cancer 1976;38:273-278. 31. Dawes C, Wood CM. The contribution of oral minor mucous gland secre- 15. Narhi TO, Meurman JH, Ainamo A et al. Association between salivary flow tions to the volume of whole saliva in man. Arch Oral Biol 1973;18:337- rate and the use of systemic medication among 76-, 81-, and 86-year-old 342. inhabitants in Helsinki, Finland. J Dent Res 1992;71:1875-1880. 32. Persson RE, Izutsu KT, Truelove EL et al. Differences in salivary flow rates 16. Osterberg T, Hedegard B, Sater G. Variation in dental health in 70-year-old in elderly subjects using xerostomatic medications. Oral Surg Oral Med Oral men and women in Goteborg, Sweden. A cross-sectional epidemiological Pathol 1991;72:42-46. study including longitudinal and cohort effects. Swed Dent J 1984;8:29-48. 33. Saunders RH, Handelman SL. Effects of hyposalivatory medications on sa- liva flow rates and dental caries in adults aged 65 and older. Spec Care Dent 17. Johnson G, Barenthin I, Westphal P. Mouthdryness among patients in long- 1992;12:116-121. term hospitals. Gerodontology 1984;3:197-203. 34. Rhodus NL, Brown J. The association of xerostomia and inadequate intake 18. Kreher JM, Graser GN, Handelman SL. The relationship of drug use to den- in older adults. J Am Diet Assoc 1990;90:1688-1692. ture function and saliva flow rate in a geriatric population. J Prosthet Dent 35. Wu AJ, Ship JA. A characterization of major salivary gland flow rates in the 1987;57:631-638. presence of medications and systemic diseases. Oral Surg Oral Med Oral 19. Sreebny LM, Schwam SS. A reference guide to drugs and dry mouth. Ger- Pathol 1993;76:301-306. odontology 1986;5:75-99. 36. Winer JA, Bahn S. Loss of teeth with antidepressant drug therapy. Arch Gen 20. Rundegren J, van Dijken J, Mornstad H et al. Oral conditions in patients Psychiatry 1967;16:239-240. receiving long-term treatment with cyclic antidepressant drugs. Swed Dent J 37. Bretz WA, Krahn DD, Drury M et al. Effects of fluoxetine on the oral envi- 1985~955-64. ronment of bulimics. Oral Microbol Immunol 1993;8:62-64. 21. Atkinson JC. Salivary gland dysfunction. Clin Geriatr Med 1992;8:499-511. 38. Navazesh M. Xerostomia in the aged. Dent Clin North Am 1989;33:75-80.